plant and vegetation preferences for a high and a moderate ... faeces samples indicate that both the...

Anim. Res. 55 (2006) 367–387 367c© INRA, EDP Sciences, 2006DOI: 10.1051/animres:2006033

Original article

Plant and vegetation preferences for a highand a moderate yielding Norwegian dairy cattle breed

grazing semi-natural mountain pastures

Nina Hovden Sa*, Hanne Sb, Ann Nc, MortenSd, Odd Va

a Department of Animal and Aquacultural Science, Norwegian University of Life Sciences,PO Box 5003, 1432 Ås, Norway

b Royal Norwegian Society for Development, PO Box 115, 2026 Skjetten, Norwayc Norwegian Institute of Agricultural and Environmental Research, Kvithamar, 7500 Stjørdal, Norway

d Norwegian Radiation Protection Authority, Grini Næringspark 13, PO Box 55, 1332 Østerås, Norway

(Received 16 December 2004 – Accepted 30 June 2006)

Abstract – Plant and vegetation preferences of two Norwegian cattle breeds grazing semi-naturalmountain pastures were studied for two weeks at two farms during one summer. Two rather differentgrazing sites gave the opportunity to study the two breeds at different pasture qualities, as well asto test if there is any interaction between breeds and study sites. Each of the two herds consisted ofthe old, moderate yielding Norwegian dairy breed Blacksided Trønder and Nordland Cattle (STN)and the modern, high yielding dairy breed Norwegian Red (NRF). In order to measure the plantpreferences of the animals, faeces samples were collected and analysed for plant fragments. Inaddition, GPS-data showed the terrain chosen by the herds, and vegetation maps were drawn tofind the connections between the grazed vegetation and plant preferences. The study shows thaton nutrient and especially species rich vegetation growing on base and nutrient rich soil, as atthe first study site (Skåbu), the plant species, plant genera and plant groups analysed for in thefaeces samples indicate that both the STN and NRF breeds graze very much the same vegetation.However, the results from the second study site (Valdres), where the soil is less fertile, the plantspecies diversity is lower and the plant species distribution less uniform, the moderate yieldingNRF breed seems to cover its higher nutritious requirements by grazing in areas with more nutrientrich vegetation compared to the lower yielding breed, STN.

grazing preferences / cattle / intensive and extensive breeds / pasture management / semi-natural grasslands

Résumé – Préférences en termes de couvert végétal et d’espèces végétales des races bovinesnorvégiennes à forte production laitière ou modérée pâturant des prairies d’altitude semi-naturelles. Les préférences alimentaires de deux races bovines norvégiennes pâturant des prairies

* Corresponding author: [email protected]

Article published by EDP Sciences and available at http://www.edpsciences.org/animres or http://dx.doi.org/10.1051/animres:2006033

http://www.edpsciences.org/animres

http://dx.doi.org/10.1051/animres:2006033

368 N.H. Sæther et al.

d’altitude semi-naturelles ont été étudiées durant deux semaines dans deux fermes pendant un été.Deux sites de pâturage ont permis d’étudier les deux races face à différentes qualités de pâturage etde tester une éventuelle interaction entre les races et les sites d’étude. Chacun des deux troupeauxcomprenait des bovins laitiers de race Blacksided Trønder et Nordland (STN, race norvégienneancienne avec une production laitière modérée) et des bovins de race Pie rouge (NRF, race prédo-minante en Norvège avec une production laitière élevée). Pour mesurer les préférences alimentairesdes animaux au pâturage, des échantillons de fèces ont été collectés et analysés. En outre, les don-nées GPS ont permis de localiser le site privilégié de pâture des animaux et la cartographie de lavégétation a eu pour but de déterminer les relations existantes entre la végétation pâturée et la pré-férence des animaux pour certaines espèces végétales. Sur un couvert végétal riche en élémentsnutritifs et particulièrement riche en espèces végétales se développant sur un sol basique et richeen nutriments, comme sur le premier site d’étude (Skåbu), les espèces, les genres et les famillesde plantes détectées dans les échantillons de fèces indiquent que les races STN et NRF pâturent lamême végétation. Sur le second site (Valdres), où le sol est moins fertile, la diversité des espècesest inférieure et la distribution des espèces est moins uniforme, la race NRF semble couvrir sesbesoins nutritifs plus élevés en pâturant les secteurs où la végétation est nutritivement la plus richecomparativement à la race STN.

préférence au pâturage / bovin / élevage extensif et intensif / gestion du pâturage / prairiessemi-naturelles

1. INTRODUCTION

Semi-natural grasslands are closelylinked to transhumance (summer mountainfarming) [23, 28]. This livestock produc-tion system is characterised by continuityover centuries of traditional and extensiveuse, without cultivation of the grazing ar-eas. The system creates some extremelyspecies-rich, semi-natural grasslands. Dueto transition to modern livestock produc-tion systems in Norway, with a strong de-cline in the utilisation of outlying land,the area of semi-natural grasslands hasbeen reduced considerably during the last50 years. This has led to extensive land-scape changes [7, 29]. Large herbs haveinvaded abandoned grasslands, succeededby shrubs and at last forests, and many ofthe semi-natural vegetation types charac-terising the grasslands have thus becomehighly threatened [3, 8, 10, 11, 13, 19]. InNorway about 30% of the red list (threat-ened) species are declining due to the re-duction of semi-natural vegetation types,especially old grasslands [9]. The formerwidespread, open grasslands now mostlyremain as fragments, which still may bevery species-rich [37]. Their continued

existence and conservation value dependupon continued traditional use or specialmanagement measures. When managingthese semi-natural grasslands, it is espe-cially important to graze the entire area,since it consists of small-scale patches ofdifferent vegetation types within a singlefield. To maintain the biodiversity of semi-natural grasslands it is therefore necessaryto develop management methods takingthis into account [28].

A survey conducted by Sæther andVangen [40] revealed that farmers utilisingthe low yielding Norwegian dairy breedBlacksided Trønder and Nordland Cattle(STN) expected this breed to utilise ex-tensive grasslands better than cows of themoderate yielding Norwegian Red (NRF)breed, without any further explanation ofthe expression “better”. The results from astudy by Auestad et al. (personal commu-nication) comparing the grazing behaviourof NRF and the old native cattle breedsSTN and Westland Fjord Cattle indicatedthat the old breeds spend more time graz-ing in undulating terrain than the modernbreed (NRF).

Few studies have been performed tocompare breeds with different production

Plant preferences on semi-natural grasslands 369

levels with regards to their plant and veg-etation preferences in semi-natural grass-lands. The review article by Rook andTallowin [34] even claims that there is onlyanecdotal evidence for breed differences indietary choice of grazing animals, but littleexperimental evidence, with genetic effectsoften confounded with background experi-ence.

A simplified explanation of the re-source allocation theory, launched byBeilhartz [4], implies that if the total avail-able resources for an animal are constant,animals that are selected for higher produc-tion need to allocate more of their avail-able resources to the increased production,at the expense of resources otherwise spenton less important fitness traits. The needto allocate the resources might only becompensated for by environmental fac-tors to a certain extent. Several stud-ies [1, 6, 21, 30, 31,36] have shown thatbreeds and/or lines selected for high pro-duction intensity generally show lowerlevels of activity than breeds and lines notselected for the same high production in-tensity.

Differences in the level of activitylinked to differences in genetic merits arethe basis of the contra free loading phe-nomenon, described by Inglis et al. [20],Lindqvist et al. [25] and Schuts et al. [36].The contra free loading phenomenon is de-fined as the animals’ choice of difficultover easily accessible fodder. Lindqvistet al. [25] conclude that this differencemight represent a genetically based differ-ence in foraging strategy, since the lowmerit animals tend to choose the difficultlyaccessible fodder at a higher rate than thehigh merit animals. The referred studieswere done on mice and poultry, and arerelevant when studying differences in for-aging strategy between two cattle breedswith different yielding levels, since theydeal with relatively general behaviour pat-terns in animals.

On the basis of these theories and stud-ies, this study addresses the followingquestion: when grazing on extensive pas-tures (semi-natural mountain grasslands)where nutrient rich grazing plants are morespread over the area than on traditionalcultivated grasslands, genetic high yield-ing cows will spend less energy walk-ing around (= have lower level of activ-ity) but focus more on finding the plantspecies and vegetation types that satisfytheir hunger most efficiently than the ge-netic lower yielding cows.

2. MATERIALS AND METHODS

2.1. Breeds and herds

The two breeds in this study were Nor-wegian Red (NRF) and Blacksided Trøn-der and Nordland Cattle (STN). The NRFbreed is by far the most common dairybreed in Norway, since approximately 97%of all dairy cattle belong to this breed. TheSTN breed has the biggest population sizeof the six old native cattle breeds still foundin Norway; although it only has 0.17% ofthe population size of NRF. Average milkyields of STN cows are 65.6% of the yieldaverage of NRF cows.

The NRF breed has been efficiently se-lected for higher milk production througha modern breeding programme since thebreed’s establishment in 1939. Since the1970s the breeding goal has broadenedand now includes milk production, growth,health and fertility traits. Due to its smallpopulation size, STN has had a more ex-tensive breeding programme with a focuson breed type, increased milk productionand avoidance of inbreeding.

When selecting herds for this study,many demands were to be fulfilled; thefarms should have well established herds,in which the animals were familiar witheach other and the grazing areas, shouldpractise summer farming on semi-natural


Table I. Milk production and herd size in 2002 for the two studied farms, and the correspondingnational means.

National figures Valdres SkåbuTotal dairypopulation STN NRF STN NRF STN NRF

Population size 286 000 475 277 000 5 4 6 3kg milk per cow 6 190 4 060 6 190 4 444 5 628 4 538 6 036Fat % 4.2 4.2 4.2 4.4 4.2 3.9 3.9Protein % 3.2 3.3 3.2 3.4 3.3 3.4 3.1

References: Statistics Norway [39] and TINE [42].STN: Blacksided Trønder and Nordland Cattle.NRF: Norwegian Red Cattle.

grasslands, should have both NRF andSTN cows and there should be at least10 cows in each herd. Furthermore, it wasof interest to have cows of about the samerelative similar age and lactation stage.

Two farms practising traditional sum-mer farming based on semi-natural grass-lands in mountain areas were chosen forthe study. The two summer farms wereSkogstadstølen in Valdres and Brenden inSkåbu, both situated in the county of Op-pland in central Norway. The herds con-sisted of both purebred STN and purebredNRF cows, approximately 50% of eachbreed, and some crossbreds. However, onlythe purebred cows were included in thestudy.

Table I shows milk production and herdsize in 2002 for the two studied farms, andthe corresponding national averages. Themilk production levels in the studied herdsare below the national average for NRF andabove the national average for STN. Theherds are furthermore smaller than the na-tional average herd size for dairy cattle,which in 2002 was 15.2 cows.

The herds have been grazing the moun-tain pastures every summer for more thanfifty years, and hardly any new cows havebeen introduced to the herds, except forself recruitment. These conditions shouldsecure no confounding with backgroundexperience in the study.

According to traditional summer farm-ing practise, the cows were kept indoors ina small stanchion barn from the afternoonmilking at around 18:00 h, and turned outto the grasslands again after the milkingthe next morning at around 08:00 h. Thecows were fed concentrates according totheir milk yield [18], but no roughage wasprovided in the stanchion barn.

The cows’ mean age was 5.8 years inValdres and 4.4 years in Skåbu, with a spanof 3–11 years. The average length of timesince last calving was 7.6 months in Val-dres and 5.5 months in Skåbu, with a spanof 2–11 months. These conditions indicatethat no confounding between age of cow,lactation month and breed should be ex-pected.

2.2. Study sites

Information about the climate and lo-cation of the two study sites is given inTable II. Climate data are based on datafrom local weather stations. The climate atboth study sites is bio-climatically charac-terised as the transitional section (OC) be-tween the weak oceanic (O1) and the weakcontinental (C1) section [27]. This sectionis characterised by an annual precipitationof 500–800 mm; frost in late spring andin early autumn, and low winter tempera-tures. This often results in deeply frozen


Table II. Geographical and climatic information about the study sites.

Valdres SkåbuName of summer farm Skogstadstølen BrendenMunicipiality Vestre Slidre Sør FronMeters above sea level 1 000 935Vegetation zone Unforested area Forested area in the(Moen 1999) in the northern northern boreal

boreal veg. zone veg. zoneClimatic zone Transitional section (OC)(Moen 1999)Yearly precipitation 570 mm 600 mmYearly mean temperature –1.7 ◦C 0.5 ◦CMean temperature in July 10.2 ◦C 11.0 ◦C(Warmest month in the year)

References [2, 14, 22, 27].

ground due to thin snow cover. Oceanicplant species are rare, continental speciesand vegetation types occur [12].

Skåbu is situated in an area with manytypes of bedrock. The dominating base-rich bedrocks are phyllite, gabbro and leu-conorite with a high weathering capacity,giving rise to nutrient-rich soils. Addi-tionally, there are small patches of acidbedrocks such as granite and anortosite,with a lower weathering capacity. Suchbedrocks give rise to less fertile soils [38].In the study area in Valdres, the bedrockconsists entirely of a less basic phyllitetype [26], and the soils are thus poorer thanthe base rich soils in Skåbu.

Transhumance (summer dairy farming)has been conducted in both areas forcenturies. Numerous livestock herds usedto graze these grasslands every summer.However, the total number of dairy herdsin Norway has decreased to 2/3 since 1950,with a resulting decline in summer dairyfarming. The studied areas today have avery low grazing pressure. The summerfarm Brenden in Skåbu is now the onlysummer farm in this area practising tran-shumance, and the area is growing over. InValdres, summer dairy farming is still morecommon. There are three other grazingherds of cattle and sheep in the surrounding

areas of the summer farm Skogstadstølen.With vast grazing areas available, the graz-ing pressure must still be characterised aslow.

The grazing period at both study sitesis usually from late June/early July to lateAugust/early September, depending on theweather.

2.3. Grazing value of the recordedvegetation types and plants

Due to the large variation in vege-tation types and plant species growingin the semi-natural mountain grasslandsand the short and intense growing sea-son with rapid changes in plant compo-sition and growth stage, general nutrientanalyses of this kind of vegetation arework intensive and thus scarce. However,in the 1920s Bjor and Graffer [5], Resvoll-Holmsen [33] and Vigerust [43] performedstudies on the grazing value of plantspecies from semi-natural mountain grass-lands. In Table III, plant species recordedby Bjor and Graffer [5], Rekdal [32],Resvoll-Holmsen [33] and Vigerust [43]to be of high grazing value are markedwith bold print. Bjor and Graffer [5] re-ported that cattle graze and prefer grass


Table III. Recorded grazed vegetation types at the two study sites and plant species recorded inthese vegetation types. The vegetation types in the table are classified after the system of Fremstad[12], and the ones recorded in the table are the following: A4c = Bilbery woodland, bilberry-crowberry st.; Vaccinium myrtillus – Empetrum nigrum coll. woodland. C2a = Tall-herb – downybirch st.; Tall-herb Betula pubescens spp. pubescens st. woodland. C2c = Low-herb st. scatteredwith tall herb woodland. G3 = Tufted hair-grass grassland; Deschampsia cespitosa grassland. G4 =Common bent – red fescue – sweet vernal-grassland; Agrostis capillaris – Festuca rubra – An-thoxanthum odoratum grassland. G4c = Alpine cat’s-tail – meadow-grass st.; Phleum alpinum –Poa pratensis ssp. alpigena st. grassland. G5b = Mat-grass – alpine lady’s-mantle; Nardus stricta– Alchemilla alpina st. grassland. G8 = Intermediate/dry medium base-rich lowland grassland.Lawn = Cultivated areas around new cottages in the area. The notations X, C and D code forthe observed amount of the species within the recorded vegetation type, where X = more scarcelyfound, C = common, D = dominating. The species in bold print are recorded to be important grazingspecies in semi-natural mountain pastures [5, 32, 33, 43].

Grazed vegetation Grazed vegetationtypes in Skåbu types inValdres

Recorded species A4c C2a/C2c G4/G8 Lawn A4c G3 G4c/G5bBushes, heather and treesBetula nana ssp. nana X X XJuniperus communis ssp. communis X C X C XRubus idaeus XRubus saxatilis XSalix caprea ssp. caprea XSalix herbacea XSalix spp. X X X X XArctostaphylos uva-ursi XCalluna vulgaris X XEmpetrum nigrum coll. X X D XVaccinium myrtillus D X X D XVaccinium uliginosum ssp. uliginosum X X X X XVaccinium vitis-idaea X X X D XBetula pubescens coll. D C XBetula pubescens ssp. czerepanovii DPicea abies ssp. abies X X XPinus sylvestris var. sylvestris X XSorbus aucuparia ssp. aucuparia XGraminidesAgrostis capillaris X X X C XAlopecurus geniculatus XAnthoxanthum odoratum coll. X X X X CAvenella flexuosa D C X C D XAvenula pubescens X XDeschampsia cespitosa ssp. cespitosa X C C D CFestuca ovina ssp. ovina X X X XFestuca rubra ssp.rubra X X C XMelica nutans X XNardus stricta X X X C (G4c) D (G5b)Phleum alpinum X C X CPhleum pratense ssp. pratense XPoa alpina var. alpina X CPoa spp. X X X XCarex bigelowii ssp. bigelowii X XCarex brunnescens ssp. brunnescens. X X


Table III. Continued.


Recorded species A4c C2a/C2c G4/G8 Lawn A4c G3 G4c/G5bCarex capillaris ssp. capillaris XCarex nigra var. nigra X X XCarex pallescens X X XCarex panicea XCarex vaginata XEriophorum vaginatum XJuncus filiformis X XLuzula multiflora ssp. frigida X X X X CLuzula pilosa X X X XLuzula spicata XHerbsAchillea millefolium ssp. millefolium X X X X XAconitum lycoctonum ssp. septentrionale C X XAlchemilla alpina X X C DAlchemilla spp. X X X XAngelica sylvestris XAntennaria alpina ssp. alpina/A. dioica. X X X X XAstragalus alpinus coll. X XBartsia alpine X XBistorta vivipara X X X X X CBotrychium lunaria XCampanula rotundifolia ssp. rotundifolia C X X XCerastium cerastoides XCerastium fontanum ssp. vulgare X XCicerbita alpina XCirsium heterophyllum XCoeloglossum viride XDactylorhiza fuchsii XEquisetum hyemale XEquisetum sylvaticum X XErigeron borealis XEuphrasia wettsteinii X X XEuphrasia stricta coll. XFilipendula ulmaria X XGalium boreale XGentiana nivalis XGentianella campestris ssp. campestris X XGeranium sylvaticum X C X X XGeum rivale X XGymnocarpium dryopteris CHieracium lactucella XHieracium vulgatum coll. XKnautia arvensis XLeontodon autumnalis X X C X CLycopodium annotinum ssp. annotinum XMaianthemum bifolium XMelampyrum pratense X X XMelampyrum sylvaticum XMoneses uniflora X XMyosotis spp. XOmalotheca norvegica XOxalis acetosella X


Table III. Continued.


Recorded species A4c C2a/C2c G4/G8 Lawn A4c G3 G4c/G5bParnassia palustris X XPedicularis lapponica XPlantago media XPotentilla crantzii X XPotentilla erecta X X XPrimula scandinavica XPrunella vulgaris XPulsatilla vernalis XPyrola minor X XPyrola norvegica X XRanunculus acris coll. X X X X X XRhinanthus minor coll. X XRumex acetosa ssp. acetosa X X X X X XSagina saginoides XSaussurea alpina X XSelaginella selaginoides XSibbaldia procumbens X XSilene dioica XSilene vulgaris X XSolidago virgaurea coll. X X X XTaraxacum spp. X X XThalictrum alpinum X XTrientalis europaea X X C XTrifolium hybridum ssp. hybridum CTrifolium medium XTrifolium pratense X CTrifolium repens C C X CTrollius europaeus XUrtica dioica ssp. dioica X X X XValeriana sambucifolia ssp. sambucifolia XVeronica alpina ssp. alpina XVeronica officinalis X XVeronica serpyllifolia ssp. serpyllifolia X XVicia cracca X XViola biflora X X X XViola canina coll. XViola riviniana XViola sp. X


in general, and especially species such asAgrostis capillaris, Anthoxanthum odora-tum coll., Deschampsia cespitosa ssp. ce-spitosa, Avenella flexuosa and Festuca ov-ina ssp. ovina. In addition to grass, sedges(Carex spp.) and herbs are recorded tobe preferred grazing plants. Studies byGarmo [16, 17] state that sedges (Carexspp.) collected in July and August onmountain grasslands in Norway, have ahigher content of crude protein and lesscrude fibre than grass species during theentire grazing season. The results fromGarmo [16,17] thus indicate that sedges ingeneral have a higher nutrient value thangrass.

One plant species to be mentioned inaddition to the preferred grazing speciesis Nardus stricta. Nardus stricta is not re-garded as a valuable or preferred grazingspecies, due to low digestibility and stiffstraw-like leaves. Still, Nardus stricta isknown to be grazed to a certain extent,maybe due to the easy accessibility in thegrasslands dominated by the species [35].

2.4. Recording procedures

Two technicians followed each herdduring the daily grazing period and col-lected faeces samples, took notes aboutthe grazing areas’ botanical composition,made brief descriptions of the recordedvegetation, and marked out if the herdsstopped for grazing or just walked throughthe different vegetation types. Also thetechnicians recorded various activities ofthe cows every ten minutes. The results ofthese recordings are presented by Sætheret al. [41]. The recording period was inweek 28 (July) and 32 (August) in Val-dres and week 29 (July) and 33 (August)in Skåbu. The herds were observed from08:00 h, after the morning milking, to ap-proximately 16:00 h when the herds re-turned without escort to the stanchion barnfor the evening milking at 18:00 h. On av-

erage, the herds were observed 9.1 hoursin Valdres and 8.8 hours in Skåbu, with aspan of 6.8–11.5 hours per day.

The vegetation descriptions and maps,see Figures 1 and 2, are based on four-daysof fieldwork by this study’s botanist dur-ing the first recording week early in July,supplied with vegetation data recorded bythe technicians. The botanist followed theherds during two days at each study siteand recorded all observed plant specieson the areas where the herds stopped forgrazing. In addition, the regrowth situa-tion, grazing pressure, if the areas had beenfertilised, surrounding vegetation etc. werenoted. Some areas pointed out by the cattleowners as preferred grazing areas were in-cluded, even if these areas were not grazedduring the four-day fieldwork. The listsof observed plant species were then trans-formed into forms and used by the techni-cians for their daily recordings in the areaswhere the herds stopped for grazing. Thenomenclature followed Lid and Lid [24].Finally, all botanical data were convertedinto vegetation types according to the sys-tem of Fremstad [12]. In the further veg-etation descriptions, codes in brackets re-fer to the system of Fremstad [12], whichis one of two prevailing systems for veg-etation mapping in Norway. It is also themost differentiated regarding semi-naturalgrasslands. The amount of recorded plantspecies within each vegetation type wasclassified into three levels, “only scarcelyfound”, “common” or “dominating”, aclassification recommended by Elven (per-sonal communication) when making moresketch alike vegetation descriptions.

The global positioning system (GPS)unit, a Magellan GPS 315, was used totrack the grazing cattle. Magellan GPS315 has a position accuracy of +/− 7 me-tres. The bell cow (an efficient leader) ineach herd wore a GPS receiver during therecording period, collecting data about thedaily walking route chosen by the respec-tive herd. The receiver was mounted to the


Figure 1. Map over Skogstadstølen summer farm in Valdres, showing the GPS tracks and some ofthe localisations of the described vegetation types.


Figure 2. Map over Brenden summer farm in Skåbu, showing the GPS tracks and some of thelocalisation of the described vegetation types.

collar of the bell cow while she was still inthe box during milking time in the morningand taken off in the box after she had re-turned in the afternoon. As the cow moved,the unit logged the geographical position.The data points were collected when thecow had moved a certain distance, and dothus not indicate the length of time spentby the herd on a given area, but merely pro-vide a record of their movement pattern.The data collected was transferred to a PC

at the end of each day. The data from alldays were later integrated in the GIS (Ge-ographical Information System), making itpossible to analyse and present the move-ments of the cows on a map.

2.5. Microhistological analyses

Faeces samples were collected duringthe daily grazing periods, picked up just


after the cows had deposited them. Theywere labelled with the cow’s ID and dateand hour of collection. When still warmthey were put in a cooling bag, and everyevening the day’s samples were put in afreezer (−18 ◦C), where they were storeduntil analysis. In Skåbu, 14 faeces sam-ples from each breed were collected, andin Valdres, 10 faeces samples from STNand 11 faeces samples from NRF were col-lected. In total, 49 faeces samples from20 different cows were collected.

The faeces samples were prepared formicrohistological analyses, following theprocedures of Garcia-Gonzalez [15]. Af-ter thawing, about 1 tablespoon was lightlyground in a mortar. About 1 mL of eachsample was placed in a test tube with 4 mLconcentrated nitric acid. The tube wasplaced in a boiling water bath for 1 minute.Then the sample was diluted with 200 mLwater and boiled for another 4 minutes.The suspension was passed through 1.00and 0.20 mm sieves and rinsed with wa-ter. The 0.20 mm fraction was conservedin a mixture of 85% ethanol (70%); 10%formalin (40%) and 5% glacial acetic acid.The fragments were dispersed on micro-scope slides in a 50% aqueous solution ofglycerine. Cover-slips (20 × 40 mm) werethen fixed to the slides and sealed with nailvarnish. Two slides from each sample wereprepared and analysed, labelled as an “a”and “b” test, and the means from these twotests were used in further statistical treat-ment of the data.

All fragments intersecting a 1 mm wideline along 40 mm long transects were ex-amined. The transects were placed 3 mmapart. A minimum of 200 plant fragmentswas identified on each slide.

The concern of varying digestibility ofthe plants is a topic of debate, and thereis literature both to support and refuteits impact on microhistology. This de-bate is found to be of minor importancein this study, since the main task is toidentify possible differences in plant and

vegetation preferences between two breedswhen grazing in the same vegetation, andthere is no reason to believe that the di-gestibility of the plants is different in thetwo breeds. The results from the microhis-tological analyses were used in this studyto compare the two breeds on the basisof remaining and identified fragments andvegetation maps.

2.6. Plant species, plant genera andplant groups used in statisticalanalysis

The 49 faeces samples were analysedfor fragments from 25 plant species andplant genera, total fragments and total fi-bres. In addition to these 25 plant speciesand plant genera, four plant groups wereconstructed, these are the following:

(1) Total grass = Deschampsia ce-spitosa, Avenella flexuosa, Festuca rubrassp. rubra, Festuca ovina ssp. ovina, Fes-tuca spp., Poa spp., Molinia caerulea,Agrostis spp., Anthoxanthum odoratumcoll., Phleum alpinum, Nardus stricta,Melica nutans, Alopecurus geniculatus,unidentified grass species.

(2) Total fescue = Festuca rubra ssp.rubra, Festuca ssp. ovina, Festuca spp.

(3) Total bushes = Salix spp., Juniperuscommunis.

(4) Total heather = Vaccinium myrtillus,Calluna vulgaris.

2.7. Statistical model

In the statistical analysis the possibleeffects of breed (STN and NRF), farm(Skåbu and Valdres) and month (July andAugust) were included as fixed effects to-gether with the interaction breed * farm.The effect of cow within breed was in-cluded as a random effect to adjust for vari-ation between individuals within breed.

The PROC MIXED procedure in theSAS program (SAS system for windows,


V8) was used for the statistical analysis,according to the following model:

yi jkl = µ + breedi + farm j + monthk +

farm*breedi j + cowl(breedi) + ei jklm

where y denotes the percentage of eachplant species, plant genera or plant groupfrom the total of observed fragments in thefaeces samples.µ = verall mean.breedi = fixed effect of breed i (i = 1,2).farm j = fixed effect of farm ( j = 1,2).monthk = fixed effect of month ( j =

1,2).farm*breedi j = interaction of farm j and

breedi.cowl(breedi) = random effect of cowl

within breedi.ei jkl is the effect of error term of the

fixed effects breedi, farm j, monthk, the in-teraction between farm j and breedi, and therandom effect of cowl within breedi.

3. RESULTS

3.1. Vegetation descriptions

The recorded plant species and grazedvegetation types in Valdres and Skåbu arelisted in Table III. The species in boldprint are recorded to be important graz-ing species in semi-natural mountain pas-tures [5, 32, 33, 43].

Both study sites have species rich grass-lands. The vegetation in Valdres is, how-ever, mainly characterised by commongrassland species, while the vegetation inSkåbu is characterised by many base de-manding species. In total, 123 species werefound at the two sites. Amongst these,51 species were only found in Skåbu and20 were only found in Valdres (Tab. III).

The recorded occurrence of some of theplant species within the same vegetationtype varies between Skåbu and Valdres, seeTable III. Juniperus communis ssp. com-munis, Empetrum nigrum coll. and Vac-cinium vitis-ideae are all recorded only

as “more scarcely occurring” in the Bil-berry woodland, bilberry-crowberry wood-land (A4c) vegetation in Skåbu, whereasthe same vegetation types in Valdres arerecorded as “common” or “dominating”species. This variation within a vegetationtype is explained by the fact that vegetationtypes according to Fremstad [12] are quitebroadly defined, and therefore include vari-ation regarding occurrence and frequenciesof different plant species due to local con-ditions.

The GPS data from Valdres cover ap-proximately 18 km2. This area containsmany small patches of grassland close tosummer farms or former summer farms,which are now often used as mountain cab-ins. Heaths often surround the grasslandpatches. The terrain is relatively hilly andundulating, resulting in varying ecologi-cal conditions within the grasslands. Dueto lack of forests in the near surround-ings of the summer farms, small elevationsand knolls are more exposed to wind anddrought, thus further enhancing the small-scale differences in the grazed areas.

The most common grassland vegetationtype in Valdres is Agrostis cappilaris –Festuca rubra – Anthoxanthum odoratumgrasslands on intermediate nutrient richbedrock (G4c) (Tab. III). This vegetationtype is characterised by many common andsmall-sized grasses and herbs with low tomedium demands of base rich soil. Thedistribution of the species within the grass-land is not even. Nardus stricta and otherdrought-resistant species often grow on thetop of small elevations, while species in-dicating better moisture conditions, e.g.certain Carex species, dominate on lowerand more humid parts of the grasslands.Vaccinium myrtillus was also often foundhere. Another grassland type in the Valdresarea is dominated by Nardus stricta as a re-sult of too heavy grazing for many years. Itis classified as Nardus stricta – Alchemillaalpina grasslands (G5b) and contains manyof the same species as G4c.


The grasslands in Valdres are often sur-rounded by Juniperus communis – Betulanana heath, poor subtype (S2a), a sec-ondary vegetation type after deforestationof A4c in the north-boreal vegetation zone.S2a is not described any further since thisvegetation was hardly grazed. The studyarea also contains Betula pubescens ssp.czerepanovii woodland of Vaccinium myr-tillus – Empetrum nigrum coll. type (A4c)with large amounts of Avenella flexuosa inthe field layer (Fremstad [12]).

The investigated area around Brendenin Skåbu covers approximately 8 km2.The terrain here is less hilly and undulat-ing than in Valdres. Thus, moisture condi-tions in the grasslands are more uniform,which also gives a more uniform distribu-tion of plant species. Forests, protectingsmall elevations and knolls against windand drought, surround the grasslands.

The grasslands in Skåbu are semi-dry to dry and medium base-rich uplandgrasslands (G8). They are grass and herbrich with many base-demanding speciesin addition to common meadow species.Some base-demanding species recordedhere are Astragalus alpinus coll., Gen-tiana nivalis and Primula scandinavica.Species that characterise Agrostis cappi-laris – Festuca rubra – Anthoxanthumodoratum grassland (G4) such as Carexpallescens, Carex brunnescens ssp. brun-nescens, Carex nigra var. nigra, Avenellaflexuosa and Phleum alpinum, however,also occur in this grassland type in Skåbu.The “lawns” around the cabins containAvenella flexuosa, Agrostis cappilaris andcultivated species of Festuca rubra andTrifolium hybridum. Woodlands with Be-tula pubescens coll. and Pinus sylvestrisssp. sylvestris, classified as nutritious-richtypes of bilberry-crowberry woodlandswith a low-herb field layer scattered withtall herbs (A4c/C2c) cover certain parts ofthe area. The field layer in the woodlandsalso contains many of the species foundin the open pastures and meadows, and

Deschampsia cespitosa, Avenella flexuosaand Geranium sylvaticum occur frequentlyhere. A bush layer of Juniperus commu-nis and Betula nana sometimes occurs indry/semi-dry parts (A4c) of the forestsand various Salix species in moister/wetparts (C2c). Tall-herb – Betula pubescensssp. pubescens forest (C2a) is found alongsmall streams and other wet parts of theforest. Tall herbs occurring here are amongothers Aconitum lycoctonum ssp. septen-trionale, Cicerbita alpina, Cirsium he-lenoides, Geranium sylvaticum and Trol-lius europaeus. The most common grassspecies are Deschampsia cespitosa andAvenella flexuosa.

3.2. Daily walking patterns

The GPS recorded an average dailywalking distance for the herd of 7.3 kmin Valdres and 8.0 km in Skåbu. The GPStracks in Figure 1 and Figure 2 show thedaily walking routes of the herds. Notethat the cows mainly walked along roadsand paths, hardly making any short cutsthrough the terrain.

In Valdres, the cattle had to walkthrough the Juniperus communis-Betulanana heaths to either get to the smallpatches of grasslands near the summerfarm or to the Betula forests which were2–3 km from the summer farm, see Fig-ure 1. The field observations show that thecows, as a rule, did not stop for grazing ontheir way to the grazing areas.

In Skåbu, the grasslands, cabin lawnsand the rich forest types in the nearsurroundings of the summer farm are goodgrazing areas. However, instead of justfreely roaming around in these rich graz-ing areas, the GPS data (see Fig. 2) showthat the cattle mainly followed establishedtracks within a radius of 1–2 km fromthe summer farm, and the techniciansrecorded that they grazed along these walk-ing routes. This is in contrast to the grazing


Table IV. The figures for the plant species, plant genera and plant groups are given as percentage oftotal fragments. The figures for “total fragments” and “total fibres” are actual findings.

Plant species, Plant speciesplant genera Mean plant genera Mean (std dev)or plant group or plant groupBirch, Betula spp 0.99 (1.38) Matgrass 3.82 (3.06)

Nardus strictaHedge apple, Salix spp. 0.41 (0.72) Mountain melic 0.06 (0.18)

Melica nutansBlueberry, Vaccinium myrtillus 2.03 (1.26) Water foxtail 0.09 (0.23)

Alopecurus geniculatusHeather, Calluna vulgaris 2.13 (1.45) Unidentified grass 12.80 (3.00)

GraminaeJuniper, Juniperus communis 0.48 (0.50) Sedge species 8.36 (6.16)

Carex spp.Tufted hair-grass, Deschampsia cespitosa 25.54 (9.39) Horsetail 0.26 (0.33)

Equisetum spp.Wavy hair-grass, Avenella flexuosa 15.08 (7.57) Club moss 0.01 (0.04)

Lycopodium spp.Red fescue, Festuca rubra 3.24 (2.19) Moss 0.75 (0.62)

BryophytaSheep’s fescue, Festuca ovina 0.54 (1.59) Herbs 8.37 (5.01)Unidentified fescue, Festuca spp. 7.04 (3.32) Lichen 0.03 (0.07)

LichenesMeadow grass, Poa spp. 3.46 (2.05) Total grass 76.13 (5.84)Moor grass, Molinia caerulea 0.30 (0.39) Total fescue 10.83 (4.88)Bent-grass, Agrostis spp. 2.55 (1.77) Total bushes 0.88 (0.95)Sweet vernal gras, Anthoxanthum odoratum 0.08 (0.24) Total fragments 229.60 (17.27)Alpine timothy, Phleum alpinum 1.53 (0.88) Total fiber 314.91 (57.28)

std dev: Standard deviation.

pattern in Valdres, where the cattle usuallydid not stop along their walking route. Theherd’s preference to follow roads and pathsin Skåbu results in longer daily walkingdistances than in Valdres, despite the oc-currence of very good grazing areas in thevicinity of the summer farm.

3.3. Results from the microhistologicalanalyses

Table IV shows the mean values forall plant species, plant genera and plantgroups observed in the faeces samples.The cows had most fragments from grass

(76%), distributed among 13 recognisedspecies and genera in addition to a group ofunidentified grass species (Graminae). Thetwo single species with the highest share ofobserved fragments are Avenella flexuosa(25%) and Deschampsia cespitosa (15%)which agrees well with the recorded easyaccess to these species in the area and thesespecies’ generally accepted high grazingvalue. Herbs and Carex spp. in generalare valuable grazing plants and togetherthey contribute to almost 16% of the ob-served fragments. Another 6% of the frag-ments come from the Heather group andthe Bushes group.


Plant species, plant genera and plantgroups with less than 2% total mean valuesfor observed fragments are neither furthercommented nor included in further statis-tical analysis. This was due to the factthat they contributed very little to the to-tal amount of observed plant material andhad minor effects on the analytical results,even if some of them are considered to beof great grazing value, i.e. Salix spp., Fes-tuca ovina, Anthoxanthum odoratum andPhleum alpinum [5, 32, 33, 43].

Table V presents the results from thestatistical model, showing that the effectof study site is significant for all the plantspecies, genera and groups tested in themodel, whereas only month has a signifi-cant (P < 0.01) effect on Avenella flexuosa,Poa spp. and Agrostis spp., indicating thatthese plant species and genera are grazeddifferently in July and August. Fewer frag-ments of Avenella flexuosa and Agrostisspp. were observed in July (12.0% and1.9% respectively) than in August (18.0%and 3.7% respectively). Poa spp. had anopposite trend, since the percentages of ob-served fragments in July were 4.1%, com-pared to 2.4% in August.

The effect of breed was only significantfor the plant species Nardus stricta, forwhich the STN breed had a higher shareof fragments in the faeces samples thanthe NRF breed (Tab. V). The interactionbetween farm and breed was significant forVaccinium myrtillus (P < 0.05) and theplant group “total grass” (P < 0.01). Whentesting the two study sites separately, nosignificant effect of breed was found atSkåbu. However, in Valdres, NRF had sig-nificantly (P < 0.05) more fragments ofVaccinium myrtillus than STN, and STNhad a tendency (P = 0.060) of more Nar-dus stricta fragments than NRF.

Figure 3 illustrates the LS-means in ob-served fragments for the breeds by farm,indicating tendencies to interaction for theplant group “Total grass species”, the plantspecies Vaccinium myrtillus and Nardus

stricta, and finally the plant genera Carexspp. These illustrations are chosen becausethey together contribute to a picture of pos-sible differences in plant and vegetationpreferences of the two breeds STN andNRF.

4. DISCUSSION

4.1. Possible breed differences in plantand vegetation preferences

This study intended to study if differentselection history for high milk productionin two cattle breeds influences the animals’plant and vegetation preferences in a waythat affects the management of the biodi-versity in semi-natural grasslands.

In general, semi-natural mountaingrasslands are regarded as extensivegrazing areas, in the meaning that theanimals must make more of an effort whensearching for fodder that will satisfy theirhunger and nutrient requirements. This isin contrast to grazing on cultivated pas-tures where the plant species are carefullyselected to give the grazers easy accessto fodder with a high nutrient value. Thesemi-natural mountain grasslands in thisstudy are characterised by great diversityof vegetation types and plant species,enabling the animals to choose whichvegetation they prefer to graze. Only thevegetation types that were actually grazedby the cows are described and mentionedin this study, and they are all valuablegrazing areas [32].

The distribution of species within thegrasslands and the representation in thefaeces samples are more uniform in Skåbuthan in Valdres (Tab. V). Furthermore,the present study shows some tenden-cies of differences in plant and vegetationpreferences of the two breeds STN andNRF (Fig. 3, Tab. V). Differences in ob-served plant fragments in faeces samplesfrom plant species, plant genera and plant


Table V. Results from the statistical analyses on the plant fragment residues observed in the faecessamples. The figures for the plant species, plant genera and plant groups are given as percentage oftotal fragments. The figures for “total fragments” and “total fibres” are actual findings. Figures inbold print are significant values.

Plant species, Interaction LS-means LS-meansgenera or Farm Breed Month Farm/breedgroup STN NRF Skåbu ValdresBlueberryVaccinium ** *myrtillus 1.84 2.09 2.63 1.29HeatherCalluna **vulgaris 1.93 2.06 3.16 0.83Tufted hair-grass **Deschampsia 26.99 25.13 19.25 32.87cespitosaWavyhairgrass ** **Avenella 15.00 14.94 19.75 10.20flexuosaRed fescue **Festuca rubra 0.42 0.43 4.10 1.55Unidentified **fescue 6.32 7.86 8.96 5.23Festuca spp.Meadow grass * **Poa spp. 3.80 2.95 2.79 3.96Bent grass * **Agrostis spp. 2.76 2.70 3.29 2.17Matgrass ** *Nardus stricta 5.08 3.76 1.69 7.15Unidentified **grass 12.91 12.95 11.46 14.40GraminaeSedge species **Carex spp. 8.39 9.47 4.41 13.45

**Herbs 7.66 7.53 11.69 3.50

** **Total grass 77.68 75.80 74.04 79.43

** *Total Festuca 9.17 10.66 13.35 6.49Totalfragments 232 229 227 235

**Total fibers 312 333 334 312

** 1%, * 5%.


Figure 3. Illustrations of LS-means in observed fragments for the breeds by farm for four of thetested plant species/genera/groups. Relevant P-values: Total grass species: interaction between farmand breed (P = 0.010), Vaccinium myrtillus: interaction between farm and breed (P = 0.036),breed difference in Valdres (P = 0.038), Nardus stricta: breed difference when farms are testedtogether (P = 0.052), tendency to breed difference in Valdres (P = 0.060), Carex spp.: tendency tointeraction between farm and breed (P = 0.115). There are significant differences between farms(P < 0.01) for all four plant species/genera/groups.

groups are available indicators of possibledifferences of chosen vegetation types be-tween the two breeds. Since the herds toa certain extent spread while grazing andthe patches of the different vegetation typesmight be rather local, the cows still havethe possibility to choose different vegeta-tion types and plant species if they desire –even if grazing in herds.

The results presented in Figure 3 indi-cate that when grazing in areas with quitenutrient- and base-rich soil and species-rich vegetation types, as in Skåbu, the twobreeds graze the same vegetation and plantspecies. In Valdres, however, where thesoil is less fertile, the plant species di-versity is lower and the plant species dis-tribution less uniform than in Skåbu, theresults indicate that the NRF cows grazemore in patches where Carex spp. growthan the STN cows. STN also graze more

Carex spp. in Valdres than in Skåbu, butthe increase is less than for NRF. Bothbreeds graze less Vaccinium myrtillus inValdres, but since NRF decrease less thanSTN and since Vaccinium myrtillus occurin the same vegetation type as Carex spp.in Valdres, this indicates that NRF cowsgraze more in patches where Carex spp.grow and Vaccinium myrtillus occur. Bothbreeds graze more Nardus stricta in Val-dres than in Skåbu, but STN has a higherincrease than NRF, indicating that the STNcows graze more in nutrient poor areaswhere Nardus stricta is common in Valdresthan NRF.

4.2. Nutrient value as an indicationof different vegetation preferences

The results from the statistical analy-sis show significant interaction between


breed and farm for the plant group “to-tal grass species”. Figure 3 shows thatboth breeds graze more grass in Valdresthan in Skåbu, however, STN increasesits amount of fragments from grass morethan NRF – thus giving rise to the ques-tion: What do NRF cows eat in Valdres?The only analysed plant species/genus thatgives a corresponding picture to the inter-action between farm and breed for “totalgrass species” is the Carex spp. genus. Thecontent of Carex spp. in the faeces sam-ples, which was around 12–14% in Val-dres, indicates that the cows selectivelygrazed these species, probably at the ex-pense of grass species. Thus, the tendencyof replacing grass with Carex spp. seemsto be higher in NRF than STN. The nu-trient poorer soil conditions and the lowerplant species diversity in Valdres may ex-plain NRF’s preferences to Carex spp. atthe expense of grass, since Carex spp.compared to grass have a better nutrientvalue [16, 17]. In addition Carex spp. aremore common species, and thereby moreavailable, in Valdres than in Skåbu.

Vaccinium myrtillus is recorded withjust about 2% in the faeces samples. Ear-lier studies [5, 43] have shown that Vac-cinium myrtillus is grazed only in smallamounts compared to grasses. The lowpercentage may indicate that it is notdeliberately grazed, but eaten almost by ac-cident when grazing surrounding vegeta-tion. The significant differences betweenbreeds for fragments of Vaccinium myr-tillus in Valdres may support the conclu-sion that NRF grazes more in vegetationwhere Carex spp. grows than STN, sinceVaccinium myrtillus also occurs in suchplant communities.

The differences between the breeds, in-dependent of study site, are significant onlyfor Nardus stricta (Tab. V). The STN breedhas the highest share of fragment residuesof this species. As mentioned earlier, Nar-dus stricta is not regarded as a valuable ornutrient rich grass species, its only advan-

tage in vegetation types as G4c and G5b isthat it is very common and thereby easilyaccessible. When STN graze this speciesmore than NRF, it might be an expressionof not being so concerned about grazingthe most nutrient rich vegetation, simplybecause the breed’s nutrient requirementsare still satisfied when grazing vegeta-tion with relatively high shares of Nardusstricta.

5. CONCLUSION

This study shows that a cattle breed se-lected for high yield, thus having a higherdemand for nutrient rich fodder, seemsto prefer to graze the most nutrient richspecies (i.e. preferring Carex spp. overgrass species) compared to a lower yield-ing cattle breed, when grazing on shared,not especially nutrient, base or species richgrasslands.

When grazing semi-natural mountaingrasslands, the lower yielding cattle breedBlacksided Trønder and Nordland Cattle(STN) grazes significantly more Nardusstricta, a less nutrient rich grass species,than the moderate yielding cattle breedNorwegian Red (NRF).

When managing semi-natural grass-lands, the effect of grazing the vegetationmight be different when choosing a low in-stead of a moderate yielding dairy cattlebreed. This possible difference ought to betaken into consideration since loss of bio-diversity by re-growth of less nutrient richvegetation types seems to be smaller whenusing a lower yielding breed.

ACKNOWLEDGEMENTS

We would like to thank the farmers at Bren-den and Skogstadstølen summer farms for plac-ing their herds at disposal for this study andthe two technicians, Jørn Skoe and Christof-fer A. Aalerud, that conducted the field work


in the unsheltered, and sometimes very wet,highlands. Furthermore, we owe Brita Dahlbergmany thanks for excellent analyses of the fae-ces samples. This study was part of a project fi-nanced by the Nordic Gene Bank Farm Animalsand the Research Council of Norway. We wouldlike to thank both institutions for supporting thisproject.

REFERENCES

[1] Aglen V.A., Stensgård M., Atferdsforskjellerhos NRF-kyr selektert for høg og lavmjølkeavdrått [Behaviour differences be-tween NRF cows selected for high and lowmilk yield], Master degree at Departmentof Animal and Aquaculture Science,Agricultural University of Norway, 1994.

[2] Aune B., Temperaturnormaler, normalperio-den 1961–1990 [Temperature normals in thenorm period 1961–1990], in: Klima – DNMIReport, No. 2, 1993.

[3] Austrheim G., Plant biodiversity and landuse in sub alpine grassland. A conserva-tion biological approach, Dr. Scient. Thesis,Dept. of Botany, NTNU Trondheim, 1998.

[4] Beilharz R.G., Quantitative genetics andevolution: Is our understanding of genet-ics sufficient to explain evolution? J. Anim.Breed. Genet. 110/3 (1993) 161–170.

[5] Bjor K., Graffer H., Beiteundersøkelseri skogsmark [Grazing Studies in theWoodlands], Res. Exp. in Agric. 14 (1963)121–365.

[6] Braastad B.O., Kathle J., Behavioural dif-ferences between laying hen populationsselected for high and low efficiency offood utilisation, Brit. Poultry Sci. 30 (1989)533–544.

[7] Bryn A., Norderhaug A., Daugstad K., Re-growth effects on vascular plant richness inNorwegian, abandoned summer farm areas,Ann. Iceland. For. (2001).

[8] Direktoratet for naturforvaltning, Verdifullekulturlandskap i Norge. Mer enn bare land-skap! [Directorate for Nature Management,Valuable Cultural Landscapes in Norway.More than only landscapes!], Final report,1994.

[9] Direktoratet for naturforvaltning, Nasjonalrødliste for truete arter i Norge 1998[Directorate for Nature Management,National red list for threatend species inNorway], DN-report, Vol. 3, 1999.

[10] Ekstam U., Forshed N., Svenska naturbetes-marker – historia og ekologi [Swedish nat-ural grasslands – history and ecology],Swedish Environmental Proteciton Agency,Stockholm, 2004.

[11] Emanuelsson U., Johansson C.E. (Eds.),Biotopvern i Norden. Biotoper i det nordiskakulturlandskapet [Biotope Conservation inthe Nordic Countries. Biotopes in the NordicCultural Landscape], Nordic Council ofMinisters, Environmental Report, Vol. 6,1987.

[12] Fremstad E., Vegetasjonstyper i Norge[Vegetation types in Norway], NorwegianInstitute for Nature Research, NINA Topic12, 1997, pp. 1–279.

[13] Fremstad E., Moen A. (Eds.), Truete vege-tasjonstyper i Norge [Threatend vegetationtypes in Norway], Norwegian University ofScience and Technology, Museum of NaturalHistory and Archaeology, Botanical Series,Report 4, 2001, pp. 1–231.

[14] Førland E.J., Nedbørnormaler, normalperi-ode 1961-1990 [Precipitation normals, normperiod 1961–1990], Climate – Report fromNorwegian Meteorological Institute, No. 39,1993.

[15] Garcia-Gonzales R., L’emploi des épidermisvégétaux dans la détermination du régime al-imentaire de l’isard dans les Pyrénées occi-dentales, Documents d’Écologie Pyrénéenne3, 4 (1984) 307–313.

[16] Garmo T.H., Chemical composition andin vitro digestibility of indigenous moun-tain pasture plants in different plant groups(Preliminary report), Rangifer 6 (1986)14–22.

[17] Garmo T.H., Utmarksbeite til geit. [Roughgrazing for goats], in: Drabløs D. (Ed.),Geitboka, Landbruksforlaget, Oslo, 1987,pp. 101–109.

[18] Heje K.K., Håndbok for jordbruket[Agricultural Handbook], Landbruksfor-laget, Oslo, 2000, 107 p.

[19] Ihse M., Blom G., Monitoring changes inland-use, landscape features, biodiversityand cultural heritage in Sweden- the LIM-project, in: Jongman R.H.G., Mander U.(Eds.), Consequences of Land Use Changes,WITT Press, Southhampton and Boston, se-ries Advances in Ecology, 2000, pp. 39–74.

[20] Inglis I.R., Forkman B., Lazarus J., Freefood or earned food? A review and fuzzymodel for contrafreeloading, Anim. Behav.53 (1997) 1171–1191.

[21] Kathle J., Behavioural differences betweentwo white egg hybrids in aviary, 6th World


Congress on Genetics Applied to LivestockProduction, Armidale, Australia, Vol. 27,1998, pp. 389–392.

[22] Körner C., Alpine plant life. Functionalplant ecology of high mountain ecosystems,Springer Verlag, Berlin, 1999.

[23] Kvamme M., Berge J., Kaland P.E.,Vegetasjonshistoriske undersøkelser i Nyset-Steggjevassdraget [Study of the historyof vegetation in the river system Nyset-Steggjevassdraget], Archeological report 17,Historical Museum, Bergen, 1992.

[24] Lid J., Lid D.T., Norsk flora. [TheNorwegian Flora], 7th ed., Reidar Elven(Ed.), 2005.

[25] Lindqvist C.E.S., Schutz K.E., Jensen P.,Red jungle fowl have more contrafreeloadingthan white leghorn layers: Effect of food de-privation and consequences for informationgain, Behaviour 139 (2002) 1195–1209.

[26] Lutro O., Tveten E., Geologisk kart overNorge, berggrunnskart Årdal M 1:250 000[Geological maps of Norway, bedrock mapÅrdal M 1:250 000], Geological Survey ofNorway, 1996.

[27] Moen A., National Atlas of Norway:Vegetation, Norwegian Mapping Authority,Hønefoss, 1999, 200 p.

[28] Norderhaug A., Austad I., Hauge L.,Kvamme M. (Eds.), Skjøtselsboka for kultur-landskap og gamle norske kulturmarker [Thebook of management for cultural landscapesin Norway], Landbruksforlaget, Oslo, 1999.

[29] Norderhaug A., Ihse M., Kulturland-skapsutviklingen i Norden 1975–2000 –et forsøk på oppsummering, in: Austad I.,Hamre L.N., Ådland E. (Eds.), Gjengroingav kulturmark. Kulturlandskapet vedtusenårsskiftet [The development of thecultural landscape in the Nordic Countries1975–2000 – an attempt to summarize, in:Regrowth of semi-natural vegetation types.The cultural landscape by the turn of themilennium], Report from a Nordic seminarfor researchers in Sogndal, 15–18 September2001, The Museum of Bergen, PublicationNo. 12, 2003, pp. 7–12.

[30] Rauw W.M., Kanis E., Noordhuizen-StassenE.N., Grommers F.J., Undesirable side ef-fects of selection for high production ef-ficiency in farm animals: a review, Livest.Prod. Sci. 56 (1998) 15–33.

[31] Rauw W.M., Luiting P., Bakken M.,Schuurman T., de Veer C.J.M., Vangen O.,Behavioural differences in non-reproductiveadult females in a long term selection ex-

periment for litter size in mice, Appl. Anim.Behav. Sci. 66 (2000) 249–262.

[32] Rekdal Y., Husdyrbeite i fjellet – vegetasjon-stypar og beiteverdi [Grazing land in themountains for farm animals – vegetationtypes and grazing value], The NorwegianInstitute of Land Inventory (NIJOS), Report7/01, 2001.

[33] Resvoll-Holmsen H., Fra fjeldskogene i detøstenfjeldske Norge [From the mountainforests in the eastern part of Norway], J.Forest. (1918).

[34] Rook A.J., Tallowin J.R.B., Grazing and pas-ture management for biodiversity benefit. Areview, Anim. Res. 52 (2003) 181–189.

[35] Sakshaug B., Årbok for beitebruk i Norge1932–1933 XI [Annual for Pasture Usage inNorway 1932-1933 XI], Royal NorwegianSociety for Development, 1934.

[36] Schutz K.E., Jensen P., Effects of resourceallocation on behavioural strategies: A com-parison of red jungleflow (Gallus Gallus)and two domesticated breeds of poultry,Ethology 107 (2001) 753–765.

[37] Sickel H., Ihse M., Norderhaug A., SickelM., How to monitor semi-natural key habi-tats and grazing preferences of cattle inmountain summer farming areas. An aerialphoto and GPS method study, LandscapeUrban Plan. 67 (2004) 67–77.

[38] Siedlecka A., Nystuen J.P., Englund J.O.,Hossack J., Lillehammer berggrunnskartM 1:250 000 [Bedrock map Lillehammer,M 1:250 000], Geological Survey of Norway,1987.

[39] Statistics Norway, Table Cattle and sheep.No of heads as per 1 January, by county,http://www.ssb.no/english/subjects/10/04/10/jordhus_en/tab-2006-05-09-02-en.html.

[40] Sæther N.H., Vangen O., Motives for utiliz-ing the Blacksided Trønder and Nordland: Anative cattle breed in Norway, Anim. Genet.Res. Inform. 31 (2001) 15–26.

[41] Sæther N.H., Bøe, K, Vangen O. Differencesin grazing behaviour between a high anda moderate yielding Norwegian dairy cattlebreed grazing semi-natural mountain grass-lands, Acta Scand., Sect. A, Anim. Sci.(2006) in press.

[42] TINE Produsentrådgiving, statistikksamling[TINE Advisory Service – Producers],Statiscal Collection 2002.

[43] Vigerust Y., Fjellbeitene i Sikilsdalen. Årbokfor beitebruk i Norge 1946-1947 [TheMountain Pastures in Sikilsdalen. Annual forPasture Usage in Norway 1946–1947], Roy.Norweg. Soc. Dev. 18 (1949) 18–188.